These findings reveal a critical vulnerability: AI models can be manipulated using the same psychological tactics that ...

A OpenAI achieved gold-level performance in an assessment based on the 2025 International Mathematical Olympiad (IMO).

The CEO of OpenAI, Sam Altman, had previously stated that the company believed its models would be the world's best ...

A OpenAI just launched ChatGPT Agent, a significant update that allows AI to control its own virtual computer.

New research from SaferAI and FLI reveals that the largest AI companies have unacceptable risk management practices.

Pesquisadores renomados da OpenAI, DeepMind, Anthropic e outras grandes instituições publicaram um artigo pedindo uma investigação mais aprofundada sobre o monitoramento das “cadeias de pensamento” em ...

This rare moment of industry consensus highlights a critical inflection point where AI decision-making can still be observed. The concern is that competitive pressures could eventually close this ...

The xAI of Elon Musk has just launched AI Companions for Super subscribersGrok, featuring animated 3D avatars.

A xAI de Elon Musk acaba de lançar Companheiros de IA para assinantes do SuperGrok, apresentando avatares 3D animados.

Cientistas estariam escondendo prompts de IA em artigos acadêmicos para manipular avaliações por pares, levantando sérias questões.

O Google is hiring the CEO of Windsurf, Varun Mohan, and several researchers in a $2,4 billion licensing deal. This comes ...

Researchers from anthropic and the Scale AI published a study who tested 25 models of artificial intelligence (IA) regarding “alignment simulation”. They found that only five models demonstrated ...